Three-dimensional surface models of the facial soft tissues acquired with a low-cost scanner

Abstract

Although there has been an increase in three-dimensional (3D) scanning methods available on the market, they are generally expensive. The DI3D system is considered a good scanner for the acquisition of soft tissue surface images. The Microsoft Kinect scanner is a much more affordable alternative for acquiring 3D models. The aim of this study was to determine whether the precision and accuracy of Kinect are similar to those of DI3D. To verify the accuracy, 10 patients were scanned with both methods The models of each patient acquired from the two scanners were superimposed using a surface-to-surface registration technique, and the distances between the models were recorded for 10 different anatomical regions of interest. For the evaluation of precision, one patient was scanned 11 different times with the Kinect scanner, and these models were compared using the same superimposition method. It was found that the average difference between the two methods was 0.3±2.03mm. The assessment of reproducibility showed an average difference between the images taken with Kinect of 0.1±0.6mm (P<0.05, one-sample t-test). Thus, Kinect showed good precision and reasonable accuracy, and appears to be an interesting and promising resource for facial analysis.